Research map / Label-efficient and robust learning / Knowledge distillation
Progressive Teacher-Student Learning for Early Action Prediction
Check your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add progressive teacher-student knowledge distillation (ablation) | STUDENT early-action model · NTU RGB-D Action | AUC (%): 59.24 -> 60.97 | Helped |
| add progressive teacher-student knowledge distillation (ablation) | STUDENT early-action model · SYSU 3DHOI | AUC (%): 71.87 -> 75.35 | Helped |
| add progressive teacher-student knowledge distillation (ablation) | STUDENT early-action model · UCF-101 | AUC (%): 88.39 -> 89.64 | Helped |
| add local MSE distillation (ablation) | STUDENT early-action model · SYSU 3DHOI | AUC (%): 71.87 -> 73.53 | Helped |
| add global MMD distillation (ablation) | STUDENT early-action model · SYSU 3DHOI | AUC (%): 71.87 -> 73.08 | Helped |
| combine local MSE and global MMD distillation (ablation) | STUDENT early-action model · SYSU 3DHOI | AUC (%): 71.87 -> 75.35 | Helped |
| add local MSE distillation (ablation) | STUDENT early-action model · UCF-101 | AUC (%): 88.39 -> 89.27 | Helped |
| add global MMD distillation (ablation) | STUDENT early-action model · UCF-101 | AUC (%): 88.39 -> 89.01 | Helped |
| combine local MSE and global MMD distillation (ablation) | STUDENT early-action model · UCF-101 | AUC (%): 88.39 -> 89.64 | Helped |
| use two-stage rather than one-stage optimization (ablation) | teacher-student model · SYSU 3DHOI | AUC (%): 73.57 -> 75.35 | Helped |
| use two-stage rather than one-stage optimization (ablation) | teacher-student model · UCF-101 | AUC (%): 89.51 -> 89.64 | Helped |
| train teacher and student jointly instead of fixing a pretrained teacher (ablation) | teacher-student model · SYSU 3DHOI | AUC (%): 75.35 -> 72.49 | Hurt |
Closest papers by meaning
- Action Knowledge Transfer for Action Prediction with Partial Videos (AAAI 2019 · similarity 0.88)
- The Wisdom of Crowds: Temporal Progressive Attention for Early Action Prediction (CVPR 2023 · similarity 0.87)
- Large-Scale Weakly-Supervised Pre-Training for Video Action Recognition (CVPR 2019 · similarity 0.87)
- STEP: Spatio-Temporal Progressive Learning for Video Action Detection (CVPR 2019 · similarity 0.86)
- From Raw Video to Pedagogical Insights: A Unified Framework for Student Behavior Analysis (AAAI 2024 · similarity 0.85)
- Stable Mean Teacher for Semi-supervised Video Action Detection (AAAI 2025 · similarity 0.85)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.